Here's another way to look a Q that I've never seen discussed
before. (There's probably a reason for that but maybe is helps
someone)
You don't have to look at a pi-net as impedance transformation.
The goal is to create enough current to feed the coax. 1500W
requires 5.5A rms and a tube doesn't produce nearly that much.
Since the feed line connects to the tank, then more than 5.5A
must be circulating in the tank for the coax to pick off. So,
design the tank to provide enough Q to get more than 5.5A
circulating.
The impedance of the Pi-net's output load capacitor forms a
voltage divider with the tank coil. The less load C, the
higher its impedance and the more of the tank's circulating
current goes into the feedline. If you replaced the output C with
a variable L, it would work very similarly...but variable L isn't
easy to make.
The standard formula, Q = sqrt( Zin / Zout ), defines the
current multiplication ratio. For example, for Zin=5000 Zin
and Zout=50, that's ratio=100 or Q=10. Impedance ratio is
voltage (or current) squared. In other words, the tank
multiplies the (RMS RF) plate current by 10 and gets it to
the coax. The circulating current divides between the load
C and the coax so the Q must go a bit higher.
It's perfectly possible to use a higher Q than the minimum.
This means more circulating current and more resistive losses.
The output load C will have a higher value (lower impedance) to
route a smaller fraction of the circulating current to the coax.
The only other reason to pick a particular Q comes from a
bit of simulation I did. At just the right Q, changing the loading
capacitor doesn't require adjustment of the tune C. I remember
using a Henry 2K on 75M that acted that way. Crank
the load up and down and the tuning remained perfect.
If anyone's interested I can drag out the work
jeff, wa1hco
----- Original Message -----
From: Phil (VA3UX) <phil@vaxxine.com>
To: <amps@contesting.com>
Sent: Wednesday, April 04, 2001 9:44 PM
Subject: Re: [AMPS] 4-1000A Project
>
> At 09:14 PM 4/4/2001 -0400, you wrote:
> >
> >
> >A Q of 12 is NOT some kind of magic number. It is a suggestion,
> >and efficiency varies only a small amount if the Q is 20 or 6 if it is
> >done with good tank components.
> >
> >12 is a number that makes people feel good, more than making
> >their amps work good. The best Q might be 6 or it might be 15...it
> >all depends.
> >
>
> "It all depends ".... on what ? Finish the thought for the benefit of the
> poor guy out there that's trying to learn.
>
> This discussion has completely overlooked the attenuation of harmonics as
a
> function of tank Q. You guys need to keep in mind that this thread
started
> with a request for advice from someone wanting to build their first 4-1000
> amp. I haven't seen all that much good advice so far; just a bunch of
> incomplete thoughts and a 4-1000 "expert" that has repeatedly "bespoken"
> himself.
>
> Phil
>
>
> >As Phil said, play with a roller and watch what happens. If you are
> >looking for a noticeable efficiency peak...it won't happen.
> >
> >What you do with the input circuit can have more to do with
> >efficiency than tank Q in many cases.
> >73, Tom W8JI
> >W8JI@contesting.com
> >
> >--
> >FAQ on WWW: http://www.contesting.com/FAQ/amps
> >Submissions: amps@contesting.com
> >Administrative requests: amps-REQUEST@contesting.com
> >Problems: owner-amps@contesting.com
> >
> >
> >
>
>
> --
> FAQ on WWW: http://www.contesting.com/FAQ/amps
> Submissions: amps@contesting.com
> Administrative requests: amps-REQUEST@contesting.com
> Problems: owner-amps@contesting.com
>
>
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|